Tabular examples

These examples explain machine learning models applied to tabular data. They are all generated from Jupyter notebooks available on GitHub.

Tree-based models

Examples demonstrating how to explain tree-based machine learning models.

• Basic SHAP Interaction Value Example in XGBoost
• Catboost tutorial
• Census income classification with LightGBM
• Census income classification with XGBoost
• Example of loading a custom tree model into SHAP
• Explaining a simple OR function
• Explaining the Loss of a Tree Model
• Fitting a Linear Simulation with XGBoost
• Force Plot Colors
• Front page example (XGBoost)
• League of Legends Win Prediction with XGBoost
• NHANES I Survival Model
• Speed comparison of gradient boosting libraries for shap values calculations
• parameters
• Python Version of Tree SHAP
• Scatter Density vs. Violin Plot
• Understanding Tree SHAP for Simple Models

Linear models

Examples demonstrating how to explain linear machine learning models.

• Explaining a model that uses standardized features
• Math behind LinearExplainer with correlation feature perturbation
• Sentiment Analysis with Logistic Regression

Neural networks

Examples demonstrating how to explain machine learning models based on neural networks.

• Census income classification with Keras

Model agnostic

Examples demonstrating how to explain arbitrary machine learning pipelines.

• Census income classification with scikit-learn
• Diabetes regression with scikit-learn
• Iris classification with scikit-learn
• SHAP Values for Multi-Output Regression Models
• Create Multi-Output Regression Model
• Get SHAP Values and Plots
• Reference
• Simple Boston Demo
• Simple Kernel SHAP
• How a squashing function can effect feature importance